The traditional television (TV) viewing experience and the process of accessing video media over the Internet have been converging for some time and are now becoming co-dependent. As a business model, both of these forms of media delivery have experimented over the years with monthly subscription fees and pay-per-view revenue models, but the advertising-supported model remains the dominant economic engine for both.
Financial support from advertisers is usually based on the number of viewers exposed to the advertisement with the advertiser being charged a fee for every thousand “impressions” or times that the advertisement is viewed; usually called “cost per thousand” or “CPM”. Generally speaking, higher CPM rates are charged for advertising that can provide additional details about those individual impressions, such as the time of day it was viewed or the market area where the viewing occurred. Associating the impressions with demographic information regarding those seeing the advertisement is even more valuable, particularly when that demographic is one that advertisers believe to offer better prospects for positive actions regarding the product or service being promoted.
For content accessed by personal computers or any other type of Internet-connected device, a viewer's Internet browsing activities may be readily detected and captured via various techniques. The most common technique is the use of a “cookie”, also known as an HTTP cookie, web cookie, or browser cookie. This is a small data file sent from the website being browsed to the user's browser. The browser then sends the cookie back to the server every time the website is re-loaded so that the server can be made aware of the user's previous activity on that server. This approach enables “shopping carts” to retain earlier, uncompleted purchases, and to pre-authenticate users so that they do not need to re-enter certain identification.
Cookies may also be used to build a history of previous browsing. Such information is beneficially used to enable the presentation of commercial offers that are more likely to be of interest to the user than arbitrary placements. For example, a user in Philadelphia who browses a search engine, such as Google, to look for say, “Hotels in Seattle,” would find that many websites browsed later would be displaying ads for Seattle travel, tours, entertainment, local attractions, and other custom-served offers. This is a result of certain data about the search activity being stored locally on the user's computer in the form of a data cookie.
Another common technique leverages the fact that most commercial web pages are not wholly self-contained. For a variety of technical and commercial reasons, many elements seen on the displayed web page are instead assembled “on the fly” by using content downloaded from many different servers, often geographically dispersed. Hence the screen location where a certain picture, animation or advertisement would be displayed is often actually blank when initially downloaded, but contains program instructions, most commonly in the HTML or JavaScript languages, that makes a request or “call” to the server where the needed content resides.
These requests typically include the IP address of the requesting computer, the time the content was requested, the type of web browser that made the request, the nature of the display it has to appear on, and other specifics. In addition to acting on the request and serving the requested content, the server can store all of this information and associate it with a unique tracking token, sometimes in the form of a browser cookie, attached to the content request.
Even where the web page does not need additional content to complete the user's viewing experience, this same technique can be used to gain insight into the actions and habits of the person browsing the site, which can then be used to personalize the types of advertising served to the user. This can be accomplished by programming web pages to request a graphic element from a particular server using an invisible (non-displaying) graphic file known as a “tracking pixel.” These are (usually) tiny image files (GIFs, JPEGs, PNGs, etc.) whose Internet address is put into web pages and other HTML documents. When the particular page containing such a tracking pixel is loaded, the web browser then sends a request, typically via the Internet, to a server at the address of the embedded web graphic. The addressed server sends the requested graphic file (e.g., a tracking pixel) and logs the event of the request for the specific graphic. These tracking pixel files are sometimes known by other names such as web bugs, transparent pixels, tracking bugs, pixel tags, web beacons or clear gifs. Regardless of what these token images are called, their function is largely the same.
In many commercial applications, an advertiser or its agency or other third-party service might decide to track impressions (as discussed above, an impression constitutes one person viewing one message) with a tracking pixel. Each time the advertisement is displayed, code in the displaying web page addresses some server, locally or across the Internet, containing the tracking pixel. The server answering the request then records information that can include the user's IP Address, Hostname, Device type, Screen Size, Operating System, Web browser, and the Date that the image was viewed.
In traditional TV viewing, commercial ratings data is typically collected and analyzed in an offline fashion by media research companies such as the Nielsen Company, using specialized equipment sometimes called a “Home Unit” that the research company has arranged to get connected to TV sets in a limited number of selected households. These devices record when the TV was tuned to a particular channel, however, there is currently an unmet need for reliable techniques to measure whether a specific video segment, (either broadcast content or advertisement) was actually watched by the viewer. Meanwhile, there is still no truly reliable process for confirming if and when broadcast content that has been recorded and stored on a DVR or the like is viewed at some later time.
Further, with existing monitoring services, such as Nielsen, there is a material delay between the time a program is broadcast and the availability of reliable, broadly-sampled information about what programming was watched in which markets and by what demographics is made available to either the content providers or advertisers. It is also a matter of significant controversy how valid the projections to the whole U.S. could be when they have been extrapolated from such a small sample of potential viewers (estimated to be approximately one out of every ten thousand households).
Consequently, the ability to accurately determine in near real-time exactly what TV program or advertisement each and every TV viewer in the U.S. is watching at any moment has long been an unmet market need. One reason this has been such a challenge is because it would require being able to identify not just what channel has been tuned to, but specifically what content is being watched, since the media actually being consumed by the viewer can include not just the scheduled programming but also regionally or locally-inserted advertisements, content that has been time-shifted, or other entertainment products.
Some attempts have been made to use audio matching technology to map what is being heard on the home TV set to a database of “audio fingerprints.” This is a process that purports to match the fingerprints to certain specific content. The speed and reliability of such technology that has been made commercially available to date has been found to have limitations. Video matching of screen images to known content is computationally more challenging than using audio but theoretically more accurate and useful. Matching the video segment being viewed to a database of samples (including those extracted only seconds previously from a live TV event) has offered a substantial technical challenge but has been effectively employed and is taught in U.S. Pat. No. 8,595,781, among others.